[PATCH v2 3/7] docs: Fix virt-aa-helper location

Michal Privoznik posted 7 patches 6 years ago
There is a newer version of this series
[PATCH v2 3/7] docs: Fix virt-aa-helper location
Posted by Michal Privoznik 6 years ago
The location of virt-aa-helper shown in the docs is incorrect.
The helper binary is installed under libexec dir.

Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
---
 docs/drvqemu.html.in | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/docs/drvqemu.html.in b/docs/drvqemu.html.in
index 87542afd27..93a7e6e7df 100644
--- a/docs/drvqemu.html.in
+++ b/docs/drvqemu.html.in
@@ -439,7 +439,7 @@ chmod o+x /path/to/directory
     <p>
       While users can define their own AppArmor profile scheme, a typical
       configuration will include a profile for <code>/usr/sbin/libvirtd</code>,
-      <code>/usr/lib/libvirt/virt-aa-helper</code> (a helper program which the
+      <code>/usr/libexec/virt-aa-helper</code> (a helper program which the
       libvirtd daemon uses instead of manipulating AppArmor directly), and
       an abstraction to be included by <code>/etc/apparmor.d/libvirt/TEMPLATE</code>
       (typically <code>/etc/apparmor.d/abstractions/libvirt-qemu</code>).
-- 
2.24.1

Re: [PATCH v2 3/7] docs: Fix virt-aa-helper location
Posted by Christian Ehrhardt 6 years ago
On Thu, Jan 30, 2020 at 8:06 AM Michal Privoznik <mprivozn@redhat.com>
wrote:

> The location of virt-aa-helper shown in the docs is incorrect.
> The helper binary is installed under libexec dir.
>
> Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
> ---
>  docs/drvqemu.html.in | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/docs/drvqemu.html.in b/docs/drvqemu.html.in
> index 87542afd27..93a7e6e7df 100644
> --- a/docs/drvqemu.html.in
> +++ b/docs/drvqemu.html.in
> @@ -439,7 +439,7 @@ chmod o+x /path/to/directory
>      <p>
>        While users can define their own AppArmor profile scheme, a typical
>        configuration will include a profile for
> <code>/usr/sbin/libvirtd</code>,
> -      <code>/usr/lib/libvirt/virt-aa-helper</code> (a helper program
> which the
> +      <code>/usr/libexec/virt-aa-helper</code> (a helper program which the
>

Don't get me wrong please the changes in this series seem correct to me
to have filenames/doc match the rest of the code.

But as it became clear in the discussion on v1 - the two Distributions that
actually run apparmor usually (Suse and Ubuntu) have the helper in
different places.
We seem to agree on the rules to list a multitude of paths.

But I wonder if in docs and makefiles this should be part of the .html.in
-> .html conversion to set the right paths for each.
Otherwise Suse and Ubuntu will have to carry reverts or modifications of
these changes forever.
For example the path above is in fact the correct path on Ubuntu - that
might also be the history how it became that way in the src.

Essentially this comes through config option --libexecdir=/usr/lib/libvirt,
so we would not even have to add a new argument for this.
What do others think about this (for all the patches of the series it
applies to)?



>        libvirtd daemon uses instead of manipulating AppArmor directly), and
>        an abstraction to be included by
> <code>/etc/apparmor.d/libvirt/TEMPLATE</code>
>        (typically <code>/etc/apparmor.d/abstractions/libvirt-qemu</code>).
> --
> 2.24.1
>
>

-- 
Christian Ehrhardt
Staff Engineer, Ubuntu Server
Canonical Ltd
Re: [PATCH v2 3/7] docs: Fix virt-aa-helper location
Posted by Michal Privoznik 6 years ago
On 1/30/20 8:39 AM, Christian Ehrhardt wrote:
> 
> 
> On Thu, Jan 30, 2020 at 8:06 AM Michal Privoznik <mprivozn@redhat.com 
> <mailto:mprivozn@redhat.com>> wrote:
> 
>     The location of virt-aa-helper shown in the docs is incorrect.
>     The helper binary is installed under libexec dir.
> 
>     Signed-off-by: Michal Privoznik <mprivozn@redhat.com
>     <mailto:mprivozn@redhat.com>>
>     ---
>       docs/drvqemu.html.in <http://drvqemu.html.in> | 2 +-
>       1 file changed, 1 insertion(+), 1 deletion(-)
> 
>     diff --git a/docs/drvqemu.html.in <http://drvqemu.html.in>
>     b/docs/drvqemu.html.in <http://drvqemu.html.in>
>     index 87542afd27..93a7e6e7df 100644
>     --- a/docs/drvqemu.html.in <http://drvqemu.html.in>
>     +++ b/docs/drvqemu.html.in <http://drvqemu.html.in>
>     @@ -439,7 +439,7 @@ chmod o+x /path/to/directory
>           <p>
>             While users can define their own AppArmor profile scheme, a
>     typical
>             configuration will include a profile for
>     <code>/usr/sbin/libvirtd</code>,
>     -      <code>/usr/lib/libvirt/virt-aa-helper</code> (a helper
>     program which the
>     +      <code>/usr/libexec/virt-aa-helper</code> (a helper program
>     which the
> 
> 
> Don't get me wrong please the changes in this series seem correct to me 
> to have filenames/doc match the rest of the code.
> 
> But as it became clear in the discussion on v1 - the two Distributions 
> that actually run apparmor usually (Suse and Ubuntu) have the helper in 
> different places.
> We seem to agree on the rules to list a multitude of paths.
> 
> But I wonder if in docs and makefiles this should be part of the 
> .html.in <http://html.in> -> .html conversion to set the right paths for 
> each.
> Otherwise Suse and Ubuntu will have to carry reverts or modifications of 
> these changes forever.
> For example the path above is in fact the correct path on Ubuntu - that 
> might also be the history how it became that way in the src.
> 
> Essentially this comes through config option 
> --libexecdir=/usr/lib/libvirt, so we would not even have to add a new 
> argument for this.
> What do others think about this (for all the patches of the series it 
> applies to)?

Aha! This is the missing part. I didn't realize that one can specify 
libexecdir independently. Alright, let me see if I can get these 
generated. In order to do that I would probably need to move existing 
files to XXX.in and have a simple sed to replace LIBEXEC with its actual 
value.

Meanwhile, I will merge the three patches you ACKed since they are 
independent of the rest.

Thanks,
Michal